INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

kson

-0.69

acle

-0.69

 Everest

-0.67

 Beef

-0.65

senal

-0.62

 Disneyland

-0.61

 Godzilla

-0.60

 Magicka

-0.60

 Harm

-0.60

 royalty

-0.59

POSITIVE LOGITS

arag

0.78

 BALL

0.70

ITNESS

0.68

aughtered

0.68

len

0.68

amus

0.68

ãģ£

0.66

oris

0.66

ratulations

0.66

aye

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.